Real options signaling games with applications to corporate finance

84 Pages·2011·0.42 MB·English

Real options signaling games with applications to
corporate finance
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation Grenadier, S. R., and A. Malenko. “Real Options Signaling
Games with Applications to Corporate Finance.” Review of
Financial Studies 24.12 (2011): 3993–4036. Web.
As Published http://dx.doi.org/10.1093/rfs/hhr071
Publisher Oxford University Press
Version Author's final manuscript
Accessed Tue Apr 09 14:37:06 EDT 2019
Citable Link http://hdl.handle.net/1721.1/75378
Terms of Use Creative Commons Attribution-Noncommercial-Share Alike 3.0
Detailed Terms http://creativecommons.org/licenses/by-nc-sa/3.0/
Real Options Signaling Games with Applications to
Corporate Finance
(cid:3)
Steven R. Grenadier Andrey Malenko
Stanford University Massachusetts Institute of Technology
June 2011
We thank the anonymous referee, Anat Admati, Felipe Aguerrevere, Geert Bekaert,
(cid:3)
Nina Boyarchenko, Cecilia Bustamante, Nadya Malenko, Ilya Strebulaev, Pietro Veronesi
(the editor), Neng Wang, Je⁄rey Zwiebel, seminar participants at Stanford University, and
participants at the 2010 Western Finance Association Annual Meeting in Victoria, BC,
the 2010 UBC Winter Finance Conference, and the 9th Trans-Atlantic Doctoral Confer-
ence at LBS for their helpful comments. Grenadier: Graduate School of Business, Stan-
ford University, 655 Knight Way, Stanford, CA, 94305, [email protected]. Malenko:
MIT Sloan School of Management, 100 Main Street, E62-619, Cambridge, MA, 02142,
[email protected].
1
EEEllleeeccctttrrrooonnniiiccc cccooopppyyy aaavvvaaaiiilllaaabbbllleee aaattt::: hhhttttttppp::://////ssssssrrrnnn...cccooommm///aaabbbssstttrrraaacccttt===111333666444888333999
Abstract
Westudygamesinwhichthedecisiontoexerciseanoptionisasignalofprivateinformation
to outsiders, whose beliefs a⁄ect the utility of the decision maker. Signaling incentives dis-
tort the timing of exercise, and the direction of distortion depends on whether the decision(cid:150)
maker(cid:146)s utility increases or decreases in outsiders(cid:146)belief about the payo⁄from exercise. In
the former case, signaling incentives erode the value of the option to wait and speed up
option exercise, while in the latter case option exercise is delayed. We demonstrate the
model(cid:146)s implications through four corporate (cid:133)nance settings: investment under managerial
myopia, venture capital grandstanding, investment under cash (cid:135)ow diversion, and product
market competition.
2
EEEllleeeccctttrrrooonnniiiccc cccooopppyyy aaavvvaaaiiilllaaabbbllleee aaattt::: hhhttttttppp::://////ssssssrrrnnn...cccooommm///aaabbbssstttrrraaacccttt===111333666444888333999
The real options approach to investment and other corporate (cid:133)nance decisions has be-
come an increasingly important area of research in (cid:133)nancial economics. The main under-
lying concept is that an investment opportunity is valuable not only because of associated
cash (cid:135)ows but also because the decision to invest can be postponed. As a result, when mak-
ing the investment decision, one must take into account both the direct costs of investment
and the indirect costs of foregoing the option to invest in the future. The applications of
the real options framework have become quite broad.1
One aspect that is typically ignored in standard models is that most real option exercise
decisions are made under asymmetric information: the decision maker is better informed
about the value of the option than outsiders. Given the importance of asymmetric infor-
mation in corporate (cid:133)nance, it is useful to understand how it a⁄ects real option exercise
decisions.2 Inthispaper, weexplorethisissuebyincorporatinginformationasymmetryinto
real options modeling. We consider a setting that is (cid:135)exible enough to handle a variety of
real world examples, characterize the e⁄ects of asymmetric information, and then illustrate
the model using four speci(cid:133)c applications: investment under managerial myopia, venture
capital grandstanding, investment under cash (cid:135)ow diversion by the manager, and product
market entry decisions by two asymmetrically informed (cid:133)rms.
In the presence of asymmetric information, the exercise strategy of a real option is an
important information transmission mechanism. Outsiders learn information about the
decision maker from observing the exercise (or lack of exercise) of the option, and thereby
change their assessment of the decision maker. In turn, because the decision maker is
3
aware of this information transmission e⁄ect, the option exercise strategy is shaped to
take advantage of it. To provide further motivation for the study, consider two examples
of option exercise decisions, where asymmetric information and signaling are likely to be
especially important.
Example 1. Delegated investment decisions in corporations. In most modern
corporations, the owners of the (cid:133)rm delegate investment decisions to the manager. There is
substantial asymmetric information: the manager is typically much better informed about
the underlying cash (cid:135)ows of the investment project than the shareholders. In this context,
themanager(cid:146)s decisionwhentoinvest transmits informationabout the project(cid:146)s net present
value (NPV). While in some agency settings the manager may want to signal higher project
values to boost her future compensation, in other agency settings the manager may want to
signal a lower project NPV to divert more value for her own private consumption. In either
setting, however, the manager will take this information transmission e⁄ect into account
when deciding when to invest.
Example 2. Exit decisions in the venture capital industry. In the venture
capital (VC) industry, there is substantial asymmetric information about the value of the
fund(cid:146)s portfolio companies, since the VC (cid:133)rm that manages the fund has a much better
information about the fund(cid:146)s portfolio companies than the fund(cid:146)s outside investors. In this
context, the (cid:133)rm(cid:146)s decision when to sell a portfolio company transmits information about
its value, and hence impacts outsiders(cid:146)inferences of the (cid:133)rm(cid:146)s investment skill. Because
investorinferencesofthe(cid:133)rm(cid:146)sinvestmentskillimpactthe(cid:133)rm(cid:146)sfuturefund-raisingability,
4
the (cid:133)rm will take this information transmission e⁄ect into account when deciding when to
sell a portfolio company.
We call such interactions real options signaling games, and study them in detail in this
paper. We begin our study with a general model of option exercise under asymmetric
information. Speci(cid:133)cally, we consider a decision maker whose payo⁄ from option exer-
cise is comprised of two components. The (cid:133)rst component is simply some fraction of the
project(cid:146)s payo⁄. The second component, which we call the belief component, depends on
outsiders(cid:146)assessment of the decision(cid:150)maker(cid:146)s type. The decision(cid:150)maker(cid:146)s type determines
the project(cid:146)s NPV and is the private information of the decision maker. Our central in-
terest is in separating equilibria (cid:150)equilibria in which the decision maker reveals her type
through the option exercise strategy.3 We characterize a separating equilibrium of the gen-
eral model, and prove that under standard regularity conditions it exists and is unique. The
equilibrium is determined by a di⁄erential equation given by local incentive compatibility.
We show that the implied option exercise behavior di⁄ers signi(cid:133)cantly from traditional
real options models. The (cid:133)rst-best (symmetric information) exercise threshold is never
an equilibrium outcome, except for the most extreme type: because the decision(cid:150)maker(cid:146)s
utility depends on outsiders(cid:146)belief about the decision(cid:150)maker(cid:146)s type, there is an incentive
to deviate from the symmetric information threshold to mimic a di⁄erent type and thereby
take advantage of outsiders(cid:146)incorrect belief. While information asymmetry distorts the
timing of option exercise, the direction of the e⁄ect is ambiguous and depends on the
nature of the interactions between the decision maker and outsiders.
5
The (cid:133)rst contribution of our paper is the characterization of the direction of distortion.
We show that the direction of distortion depends on a simple and intuitive characteristic:
the derivative of the decision(cid:150)maker(cid:146)s payo⁄ with respect to the belief of outsiders about
the decision(cid:150)maker(cid:146)s type. If the decision maker bene(cid:133)ts from outsiders believing that
the project(cid:146)s value is higher than in reality, then signaling incentives lead to earlier option
exercise than in the case of symmetric information. In contrast, if the decision maker
bene(cid:133)ts from outsiders believing that the project(cid:146)s value is lower than in reality, then
the option is exercised later than in the case of symmetric information. The intuition
underlying this result comes from the fact that earlier exercise is a signal of the better
quality of the project. For example, other things equal, an oil-producing (cid:133)rm decides to
drill an oil well at a lower oil price threshold when it believes that the quality of the oil
well is higher. Because of this, if the decision maker bene(cid:133)ts from outsiders believing
that the project(cid:146)s quality is higher (lower) than in reality, she has incentives to deviate
from the (cid:133)rst-best exercise threshold by exercising the option marginally earlier (later) and
attempting to fool the market into believing that the project(cid:146)s quality is higher (lower)
than in reality. In equilibrium, the exercise threshold will be lowered (raised) to the point
at which the decision(cid:150)maker(cid:146)s marginal costs of ine¢ ciently early (late) exercise exactly
o⁄set her marginal bene(cid:133)ts from fooling outsiders. Importantly, outsiders are rational.
They are aware that the decision maker shapes the exercise strategy to a⁄ect their belief.
As a result, in equilibrium outsiders always correctly infer the private information of the
agent. However, even though the private information is always revealed in equilibrium,
6
signal-jamming occurs: the exercise thresholds of all types, except for the most extreme
type, are di⁄erent from the (cid:133)rst-best case and are such that no type has an incentive to
fool outsiders.
The second contribution of our paper is illustrating the general model with four cor-
porate (cid:133)nance applications that put additional structure on the belief component of the
decision(cid:150)maker(cid:146)s payo⁄: investment under managerial myopia, venture capital grandstand-
ing, investment under cash (cid:135)ow diversion by the manager, and product market entry deci-
sions by two asymmetrically informed (cid:133)rms. The (cid:133)rst application we consider is a timing
analog to the myopia model of Stein (1989). We consider a public corporation, in which
the investment decision is delegated to a manager, who has superior information about
the project(cid:146)s NPV. As in Stein (1989), the manager is myopic in that she cares not only
about the long-term performance of the company but also about the short-term stock price.
The timing of investment reveals the manager(cid:146)s private information about the project and
thereby a⁄ects the stock price. As a result, the manager invests ine¢ ciently by exercising
her investment option too early in an attempt to fool the market into overestimating the
project(cid:146)s NPV and thereby in(cid:135)ating the current stock price.
The second application deals with the VC industry. As discussed in Gompers (1996),
younger VC (cid:133)rms often take companies public earlier than older VC (cid:133)rms to establish a
reputation and successfully raise capital for new funds. Gompers terms this phenomenon
(cid:147)grandstanding(cid:148)and suggests that inexperienced VC (cid:133)rms employ early timing of initial
public o⁄erings (IPOs) as a signal of their ability to form higher-quality portfolios. We
7
formalize this idea in a two-stage model of VC investment. An inexperienced VC (cid:133)rm
investslimitedpartners(cid:146)moneyinthe(cid:133)rstroundandthendecideswhentotakeitsportfolio
companypublic. Limitedpartnersupdatetheirestimateofthegeneralpartner(cid:146)sinvestment-
picking ability by observing when the decision to take the portfolio company public is made
and use this estimate when deciding how much to invest in the second round. Because the
amount of second round (cid:133)nancing is positively related to the limited partners(cid:146)estimate of
thegeneralpartner(cid:146)sability, thegeneralpartnerhasanincentivetofoolthelimitedpartners
into believing that her ability is higher. Since an earlier IPO is a signal of better quality
of the inexperienced general partner, signaling incentives lead to earlier than optimal exit
timing of inexperienced general partners, consistent with the grandstanding phenomenon
of Gompers (1996).
The other two applications belong to the case of the decision maker bene(cid:133)ting more
when outsiders believe that the project(cid:146)s NPV is lower than in reality, and thus imply an
ine¢ ciently delayed option exercise. Similar to the (cid:133)rst application, the third application
studies a delegated investment decision in a corporation. However, unlike the second appli-
cation, the nature of the agency con(cid:135)ict is di⁄erent. Speci(cid:133)cally, we consider a setting in
which a manager can divert a portion of the project(cid:146)s cash (cid:135)ows for private consumption,
which makes the problem a timing analogue of the literature on agency, asymmetric infor-
mation, and capital budgeting (e.g., Harris, Kriebel, and Raviv 1982; Stulz 1990; Bernardo,
Cai, and Luo 2001). In this application, private information gives the manager an incentive
to delay investment so that outside investors underestimate the true NPV of the project,
8
which allows the manager to divert more without being caught. This creates incentives to
fool outside investors by investing as if the project was worse than in reality and thereby
leads to later investment than in the case of symmetric information. In equilibrium outside
investorscorrectlyinfertheNPVoftheproject, butstillsignal-jammingoccurs: investment
is ine¢ ciently delayed to prevent the manager from fooling outside investors.
Finally, the fourth application we consider is sequential entry into a product market
in the duopoly framework outlined in Chapter 9 of Dixit and Pindyck (1994). The major
distinction of our application is that we relax the assumption that both (cid:133)rms observe the
potential NPV from launching the new product. Instead, we assume that the two (cid:133)rms are
asymmetrically informed: one (cid:133)rm knows the project(cid:146)s NPV, while the other learns it from
observing the investment (or lack of investment) of the better informed (cid:133)rm. As a result,
the better informed (cid:133)rm has an incentive to delay investment to signal that the quality of
the project is worse than in reality and thereby delay the entry of its competitor and enjoy
monopoly power for a longer period of time. Thus, the timing of investment is ine¢ ciently
delayed. However, the under-informed (cid:133)rm rationally anticipates the delay of investment
by the better informed (cid:133)rm, so in equilibrium the timing of investment reveals the NPV of
the product truthfully.
Our (cid:133)ndings have a number of implications. First, the e⁄ect of information asymmetry
on investment timing is far from straightforward. In fact, information asymmetry can both
speed up and delay investment, thus leading to overinvestment and underinvestment, re-
spectively. The direction of distortion depends on the nature of the agency con(cid:135)ict between
9